Qudelix-5K Bluetooth DAC & Headphone Amp

[Cahudson42]

This happens every time I re-connect to 5k. Restarting the app solves this and all custom profile names are visible again.

Thanks, @Baphomet I think my custom names came back on the second restart after FW and App upgrade - but they are indeed back . Thanks again..

 

[bobbooo]

@bobbooo
With firmware v1.5, it appears the 12 db max has been removed. I was able to enter 24 in all the filter types. Each time it 'took' and the graph reflected it.

Qudelix support just said it should be max +/-30 dB gain for each filter now by the way.

 

[Rottmannash]

Hopefully, the Qudelix 5k shows the future of $100-$200 Desktop DAC/Amps. Topping, Schitt, JDS, SMSL, Geshelli et al only need to provide PEQ, reconstruction filter selection, balanced/unbalanced, sample rate selection, digital volume control, etc. and Android/IOS Apps to control via simultaneous bt.

Do this while retaining their current 110-120 SINADs and 1 watt plus outputs, and they will have obsoleted their current products and created a nice new market for themselves...

Perhaps it will also provide a wake-up call to miniDSP to finally provide Android/IOS Apps for their devices as well. Their continuing to require a PC to off-line download control parameters reeks obsolescence, and leaves them wide open for a new competitor in their DSP space..

Would having to power this with an external power supply degrade its performance?

 

[Michael YYZ]

Is it worth switching to a Qudelix 5K from a Radsone ES100 (not MKII)? I’ve been pretty happy with the latter but the company, Radsone, seems to have gone in hibernation (e.g. no new firmware updates for almost two years). I am using the device with a pair of Shure SE846 IEMs with signal from Apple devices (AAC).

@amirm Speaking of Dhure, any chance to review the RMC-TW1 (https://www.amazon.ca/gp/product/B084SV9PGK/)? It looks like a very interesting product for my IEMs.

 

[wslee]

A big factor was that aptX LL requires a separate antenna. Manufacturers don't want to add an extra antenna only for LL.

​

https://en.wikipedia.org/wiki/AptX#aptX_LL​

AptX Low Latency requires a dedicated, wireless antenna, so it did not achieve much adoption in smartphones and was retired by Qualcomm in favor of aptX Adaptive​

No, it's not true.
Every Bluetooth Audio Codec is through the same A2DP and the same Bluetooth Classic BR/EDR link.
It's a matter of just money.
We need to pay for each codec's royalty.
Very few users need aptX-LL.
And those users don't care for the sound quality.
That's why we omitted the aptX-LL.

 

[Erispedia]

Is it worth switching to a Qudelix 5K from a Radsone ES100 (not MKII)? I’ve been pretty happy with the latter but the company, Radsone, seems to have gone in hibernation (e.g. no new firmware updates for almost two years). I am using the device with a pair of Shure SE846 IEMs with signal from Apple devices (AAC).

If you have no need for PEQ, save your money. I have ES100 and Q5s. Ordered 5K just for the PEQ.

 

[Navid]

Is it worth switching to a Qudelix 5K from a Radsone ES100 (not MKII)? I’ve been pretty happy with the latter but the company, Radsone, seems to have gone in hibernation (e.g. no new firmware updates for almost two years). I am using the device with a pair of Shure SE846 IEMs with signal from Apple devices (AAC).

@amirm Speaking of Dhure, any chance to review the RMC-TW1 (https://www.amazon.ca/gp/product/B084SV9PGK/)? It looks like a very interesting product for my IEMs.

this is what @amirm said when I asked him:
Navid said:
So, In your understanding and considering my setup, is there a benefit in getting the 5k instead of ES100? Would you pls elaborate ? (I have a HiFiMan HE4XX and using my Galaxy Note 9 to drive them)

@amirm said:
Strictly speaking, no. They are both powerful enough to drive your headphone well. That said, I always like to go for the latest version of things and the 5K is it. It has even more power just in case you get a less sensitive headphone. Its buttons are easier to press and see.

 

[bobbooo]

Is it worth switching to a Qudelix 5K from a Radsone ES100 (not MKII)? I’ve been pretty happy with the latter but the company, Radsone, seems to have gone in hibernation (e.g. no new firmware updates for almost two years). I am using the device with a pair of Shure SE846 IEMs with signal from Apple devices (AAC).

Yes it's worth it, for the parametric hardware equalizer alone, allowing for system-wide EQing using for example these settings for your Shure SE846.

 

[bobbooo]

@bobbooo
With firmware v1.5, it appears the 12 db max has been removed. I was able to enter 24 in all the filter types. Each time it 'took' and the graph reflected it.

From the Qudelix forum, apparently there's a new bug, in which filter gain values over +/- 12.8 dB won't work correctly. Could you confirm this by trying for example -14 dB for one of the filters and see if it sticks after saving the whole PEQ profile? Maybe 24 dB works but intermediate values don't.

 

[bobbooo]

One thing I've been wondering is what determines the 'width' of a device's 1 kHz fundamental peak in the FFT plots, and if this could in any way have an audible effect. For example, the Qudelix 5K's seems wider than most:

Compare with the narrower 1 kHz peak of the Samsung Galaxy S8+:

 

[Soniclife]

One thing I've been wondering is what determines the 'width' of a device's 1 kHz fundamental peak in the FFT plots, and if this could in any way have an audible effect. For example, the Qudelix 5K's seems wider than most:

Compare with the narrower 1 kHz peak of the Samsung Galaxy S8+:

Look at the width at -130, they look similar to my eye, the extra noise on the S8 hides the lower skirt shown in the 5K.

 

[bobbooo]

Look at the width at -130, they look similar to my eye, the extra noise on the S8 hides the lower skirt shown in the 5K.

While true, compare the width further up at -120 or -110 dB, and the 5K's peak is definitely wider. Then there are other devices with even narrower peaks such as the Dense Audio Adapt Reference (which uses the same ES9218P DAC/amp as the Qudelix 5K by the way):

 

[Veri]

While true, compare the width further up at -120 or -110 dB, and the 5K's peak is definitely wider. Then there are other devices with even narrower peaks such as the Dense Audio Adapt Reference (which uses the same ES9218P DAC/amp as the Qudelix 5K by the way):

Short answer, it has to do with the master clock used. The wider skirt should not result in any audible issues, in any case.

 

[Eskil]

This was Qudelix's response when I asked on their forum what the output impedance of the 5K is:


Hmm...the only reason I can think of why they would flat out refuse to publish a figure is that it isn't very low. Without knowing that figure there's a risk this device could have high output impedance and so adversely affect the frequency response of headphones/IEMs with large impedance swings.

I don't know if this has been replied to.
I have measured the output impedance in single ended mode/1V 44.1kHz mode at 1kHz. 0dB signal. I got 1.016V out unloaded and 1.011V out 33ohm loaded. 4 point measurement at the connector, measured by APx555. So at 1kHz the output impedance is a factor 200 under my 33R load resistor, so something around 170mohm. Very respectable result. The balanced output impedance is thus expected to have less than the double impedance.
My THD+n measurement at 1kHz -1dB (1V mode) gave 0.0025% both unloaded and 33ohm loaded, unchanged harmonics composition. THD+n was even the same at 20Hz, and also loaded.
I also measured Balanced out, 2V mode. Here THD+n actually dropped when 33ohm loaded. The measurements here were a bit lower.

My conclusion is that
1: Output impedance is low
2: THD+n is not badly effected by load - so the THD/sound issues don't come from weak internal supplies and lo-fi buffer. At least not in the case of stationary signals. So maybe we're talking bad ref for the DAC?
3: THD+n is still excellent at 20Hz. No bad DC blocking cap in the signal path. Peak currents in the power supply are bigger here where decoupling doesn't work any more, but it doesn't seem like a problem.
4: It doesn't seem to be weak on power - I could fear that on a probably charge pump negative supply based system.
5: The with big margin dominant FFT pins are harmonics, not spurious' from bad PCB layout.

I hear some modulation by heavy bass on voices, I hear some "roughness", I hear some "missing calm" compared to good desktop systems.
I plan to do some IMD measurements, both SMPTE type (how does a strong low freq modulate a weaker higher freq tone), and some "two similar tones at high freq FFT IMD"
Then, as my 555 doesn't have USB, my source is a PC with generated 24bit sines. I need to measure that sine in the digital domain to check its THD.
Edit:
My objective is to find where to modify it to improve sound - if easily done. Anyone has a datasheet of the 18p?

 

[bobbooo]

I don't know if this has been replied to.
I have measured the output impedance in single ended mode/1V 44.1kHz mode at 1kHz. 0dB signal. I got 1.016V out unloaded and 1.011V out 33ohm loaded. 4 point measurement at the connector, measured by APx555. So at 1kHz the output impedance is a factor 200 under my 33R load resistor, so something around 170mohm. Very respectable result. The balanced output impedance is thus expected to have less than the double impedance.
My THD+n measurement at 1kHz -1dB (1V mode) gave 0.0025% both unloaded and 33ohm loaded, unchanged harmonics composition. THD+n was even the same at 20Hz, and also loaded.
I also measured Balanced out, 2V mode. Here THD+n actually dropped when 33ohm loaded. The measurements here were a bit lower.

My conclusion is that
1: Output impedance is low
2: THD+n is not badly effected by load - so the THD/sound issues don't come from weak internal supplies and lo-fi buffer. At least not in the case of stationary signals. So maybe we're talking bad ref for the DAC?
3: THD+n is still excellent at 20Hz. No bad DC blocking cap in the signal path. Peak currents in the power supply are bigger here where decoupling doesn't work any more, but it doesn't seem like a problem.
4: It doesn't seem to be weak on power - I could fear that on a probably charge pump negative supply based system.
5: The with big margin dominant FFT pins are harmonics, not spurious' from bad PCB layout.

I hear some modulation by heavy bass on voices, I hear some "roughness", I hear some "missing calm" compared to good desktop systems.
I plan to do some IMD measurements, both SMPTE type (how does a strong low freq modulate a weaker higher freq tone), and some "two similar tones at high freq FFT IMD"
Then, as my 555 doesn't have USB, my source is a PC with generated 24bit sines. I need to measure that sine in the digital domain to check its THD.
Edit:
My objective is to find where to modify it to improve sound - if easily done. Anyone has a datasheet of the 18p?

Thanks, that confirms the low output impedance measured on here and elsewhere of ~0.2-0.4 ohms. I'd be very interested in seeing your full measurements of the Qudelix 5K using your APx555, including the SMPTE and similar-tone IMD data, and whether you can find any measured performance shortcomings that would correlate with your (presumably sighted and not level-matched) subjective impression of 'modulation', 'roughness' or 'missing calm' compared to a good desktop system (maybe including measurements of the latter for reference).

 

[Eskil]

Thanks, that confirms the low output impedance measured on here and elsewhere of ~0.2-0.4 ohms. I'd be very interested in seeing your full measurements of the Qudelix 5K using your APx555, including the SMPTE and similar-tone IMD data, and whether you can find any measured performance shortcomings that would correlate with your (presumably sighted and not level-matched) subjective impression of 'modulation', 'roughness' or 'missing calm' compared to a good desktop system (maybe including measurements of the latter for reference).

Ah, I didn't see the Ro/Zo measurement here. The totally subjective list of listening shortcomings is not level matched, ABX'ed or blind tested, my abx box is for line level and power amps. But my experience is that when THD is at that level, performance is still blind recognized.
I'll have to look up how AP recommends to measure USB devices. I think there has been some newsletters on that through the years. I normally make things where an input can be accessed.
In case someone can point me to (link) how to avoid the unlinearity in windows sound system when sound level is over -1dB, that would be helpful - though it would probably belong in another forum.

 

[bobbooo]

But my experience is that when THD is at that level, performance is still blind recognized.

That's quite the claim, to be able to hear 0.0025% distortion...Can you post the log including checksum of a successful passed test (at least 13/16 trials) using Foobar2000's ABX Comparator component? You can use this software by @pkane to generate the distorted test file to ABX with the original: https://distortaudio.org.

 

[Eskil]

You are right, it is difficult to hear THD's at that level, especially if it isn't accompanied with "something else", like inharmonic products, change in output impedance. Though this is a nice product, it does rather recognizably, at uncalibrated volume sound less perfect than the 30x as expensive desktop products that made me buy my headphones recently. I guess this is what I'm referring to. I expect to get time to measure some more next week.

 

[amirm]

I plan to do some IMD measurements, both SMPTE type (how does a strong low freq modulate a weaker higher freq tone)

That is what this graph was in the review:

 

[bobbooo]

You are right, it is difficult to hear THD's at that level, especially if it isn't accompanied with "something else", like inharmonic products, change in output impedance.

The Qudelix 5K's measured inharmonic distortion is below -130 dB (0.00003%). Are you now claiming to be able to hear that too? What do you mean by "change in output impedance"? Change with what? It's been measured at a very low 0.2 ohms. Even if it 'changed' somehow by an order of magnitude to 2 ohms under some specific conditions, this would still not have an audible effect for pretty much any headphone.

Though this is a nice product, it does rather recognizably, at uncalibrated volume sound less perfect than the 30x as expensive desktop products that made me buy my headphones recently.

If you're making all these claims of audible inferiority to "expensive desktop products", you need to provide evidence for this, blind and level-matched. The fact that the only reason you seem to be citing is the Qudelix 5K's lower price suggests that you're suffering from pricing bias and so its supposed audible shortcomings you think you're hearing are in fact illusory.